AMS (Agent Management System) is a comprehensive B2B SaaS platform designed to be the definitive command and control layer for enterprise AI agent fleet management. This repository contains both the AMS platform codebase and a sophisticated collection of 134+ specialized AI development agents for collaborative software development.
- 🏢 AMS Platform: Enterprise agent fleet management and observability platform
- 🤖 AI Agent Collection: 134+ specialized development agents with multi-agent orchestration
/backend # FastAPI backend service
/api # API endpoints and routing
/core # Core business logic
/models # Database models and schemas
/services # Business services and integrations
/tests # Backend tests
/frontend # React TypeScript frontend
/src # Source code
/public # Static assets
/infra # Infrastructure and deployment
/docker # Docker configurations
/k8s # Kubernetes manifests
/terraform # Infrastructure as Code
/docs # Documentation
/api # API documentation
/architecture # System architecture docs
/deployment # Deployment guides
/Open-SWE-With-Agents/ # 134+ AI Development Agents Collection
/development/ # 24 Development & Architecture agents
/quality/ # 15 Quality Assurance & Testing agents
/devops/ # 13 DevOps & Infrastructure agents
/product/ # 9 Product & Business agents
/design/ # 7 Design & User Experience agents
/marketing/ # 7 Marketing & Growth agents
/documentation/ # 6 Documentation & Communication agents
/orchestration/ # 5 Orchestration & Management agents
/operations/ # 5 Operations agents
/data-ai/ # 4 Data & AI Engineering agents
/security/ # 4 Security & Compliance agents
/specialized/ # 9 Specialized & Utility agents
/.taskmaster/ # Task Master AI project management
Prerequisites
- Node.js 18+
- Python 3.11+
- Docker and Docker Compose
- PostgreSQL 15+
- Redis 7+
Quick Start
# Clone the repository git clone https://github.com/jaydubya818/Agent_Management_System.git cd Agent_Management_System # Start development environment docker-compose up -d # Backend setupcd backend pip install -r requirements.txt python main.py # Frontend setup (new terminal)cd frontend npm install npm run dev
Environment Variables Copy
.env.exampleto.envand configure:# Database DATABASE_URL=postgresql://user:pass@localhost:5432/ams_dev # Redis REDIS_URL=redis://localhost:6379 # JWT JWT_SECRET_KEY=your-secret-key # External APIs ANTHROPIC_API_KEY=your-anthropic-key OPENAI_API_KEY=your-openai-key
- Fleet Command Dashboard with agent heatmap
- Distributed Trace Explorer for agent execution paths
- Real-time alerting engine with webhook integration
- AI briefings for daily/weekly fleet summaries
- First-class agent identity and profile management
- Champion/Challenger A/B testing framework
- Automated canary deployments and rollbacks
- Agent lifecycle automation and deprecation recommendations
- Policy engine for spend caps and action controls
- Immutable audit trail for all agent actions
- Role-based access control (RBAC)
- Emergency stop functionality for fleet-wide control
- Meta-Agent for autonomous cost optimization
- Model performance analysis and recommendations
- Self-healing capabilities for common errors
- Automated A/B testing for cost optimization
- Python SDK for LangGraph agent integration
- OAuth-based integrations (Jira, GitHub, Slack)
- Public REST API for programmatic control
- Comprehensive documentation and examples
- Backend: FastAPI, SQLAlchemy, Celery, Redis
- Frontend: React, TypeScript, Tailwind CSS, Chart.js
- Database: PostgreSQL, ClickHouse (events/logs)
- Infrastructure: Docker, Kubernetes, AWS/GCP
- Observability: OpenTelemetry, LangSmith
- Agent Framework: LangGraph
This project uses Task Master AI for project management:
# View current tasks
tm list
# See next task to work on
tm next
# View specific task details
tm show <task-id># Update task progress
tm update-subtask --id=<subtask-id> --prompt="Progress update"# Mark tasks complete
tm set-status --id=<task-id> --status=done- Check the current sprint tasks with
tm list - Pick up the next available task with
tm next - Create a feature branch from main
- Implement the task following the detailed requirements
- Update task progress using Task Master AI
- Submit a pull request with comprehensive testing
This project is licensed under the MIT License - see the LICENSE file for details.
This repository includes a comprehensive collection of 134+ specialized AI agents designed for collaborative software development. Each agent is a domain specialist with standardized communication protocols, enabling seamless collaboration across complex development workflows.
| Category | Count | Purpose | Key Agents |
|---|---|---|---|
| 🎯 Orchestration & Management | 5 | Project coordination and multi-agent management | agent-organizer, context-manager, tech-lead-orchestrator |
| 🏗️ Development & Architecture | 24 | Software development, frameworks, and system design | react-pro, backend-architect, python-pro, nextjs-pro |
| 🎨 Design & User Experience | 7 | UI/UX design, visual systems, and user research | ui-designer, ux-designer, brand-guardian |
| 🔧 Quality Assurance & Testing | 15 | Code review, testing, performance optimization | code-reviewer, test-automator, qa-expert |
| 🔒 Security & Compliance | 4 | Security auditing, compliance, incident response | security-auditor, incident-responder |
| 📊 Data & AI Engineering | 4 | Data pipelines, machine learning, AI systems | data-engineer, ml-engineer, prompt-engineer |
| ☁️ DevOps & Infrastructure | 13 | Cloud architecture, deployment, infrastructure | cloud-architect, deployment-engineer |
| 📝 Documentation & Communication | 6 | Technical writing, API docs, content creation | api-documenter, documentation-expert |
| 🎯 Product & Business | 9 | Product management, sprint coordination, analytics | ai-scrum-master, product-manager |
| 📈 Marketing & Growth | 7 | Growth hacking, social media, content strategy | growth-hacker, marketing-writer |
| 🏢 Operations | 5 | Analytics, finance, legal compliance, support | analytics-reporter, finance-tracker |
| 🎭 Specialized & Utility | 9 | Specialized tools, workflow optimization | workflow-optimizer, tool-evaluator |
agent-organizer: Master orchestrator for complex, multi-agent tasks with intelligent delegationcontext-manager: Central nervous system for agent coordination and project state awarenessai-scrum-master: Automated Scrum Master with 3-hour standup cycles and continuous sprint management
react-pro: Expert React developer with modern patterns, performance optimization, and testingbackend-architect: System design, API architecture, database design, and scalabilitypython-pro: Expert Python developer for backend and data applications with clean architecturenextjs-pro: Next.js specialist for full-stack React applications with SSR/SSG optimization
security-auditor: Senior application security auditor with penetration testing capabilitiescode-reviewer: Expert code review specialist with quality assessment and security reviewperformance-engineer: Application performance optimization with bottleneck identification
cloud-architect: Multi-cloud architecture with cost optimization and security designdeployment-engineer: CI/CD pipeline and deployment automation with Kubernetes orchestration
All agents follow a standardized three-phase communication protocol:
- Context Acquisition: Query
context-managerfor project state - Solution Implementation: Execute specialized tasks
- Activity Reporting: Report completion back to context-manager
@react-pro Create a responsive dashboard component with real-time data
@security-auditor Audit authentication system for vulnerabilities
@cloud-architect Design scalable AWS infrastructure for microservices@agent-organizer Build a complete e-commerce platform with payment integration
# Automatically selects and coordinates: backend-architect, react-pro, security-auditor, api-documenter@ai-scrum-master Set up automated sprint management for development team
# Establishes 3-hour standup cycles with sub-agent coordination# Navigate to agents directorycd Open-SWE-With-Agents
# Run setup script
./setup_claude_agents.sh
# Optimize agents
./optimize_agents.sh
# Validate setup
./.claude/validate_agents.sh- MCP Servers: context7, magic, sequential-thinking, playwright
- Development Tools: Git, Docker, testing frameworks
- Cloud Platforms: AWS, Azure, GCP
- Communication: Slack, Teams, Discord
For detailed agent specifications and capabilities, see Open-SWE-With-Agents/AGENTS_DETAILED_README.md.
- Project Lead: Your Name
- GitHub: Agent Management System
- Task Management: Powered by Task Master AI
- AI Agents: 134+ specialized development agents included